JP6714714B2 - Vehicle control device - Google Patents

Description

The present invention relates to a vehicle control device suitable for application to a vehicle capable of automatic driving (including automatic driving support).

Japanese Patent No. 5306934 (hereinafter referred to as JP5306934B) is provided with a plurality of modules (referred to as action purpose generation modules) that are layered according to the length of the operation cycle in order to generate a plurality of action purposes. A control system is disclosed which controls a control target through a control module according to the calculation results of a plurality of action purpose generation modules. A specific control target of this control system is a legged robot.

This JP5306934B has an action purpose generation module divided into three layers according to the operation cycle, and the action purpose generation module having the longest operation cycle is in charge of moving the robot to the target position. The intermediate action purpose generation module is in charge of preventing the robot from contacting an object, and the action purpose generation module having a short calculation cycle is in charge of stabilizing the posture of the robot.

In this hierarchical control system, the operation of the controlled object is controlled by preferentially reflecting the evaluation result by the action purpose generation module having a shorter operation cycle than the evaluation result by the action purpose generation module having a long operation cycle. (Claim 1 of JP5306934B).

By the way, in a vehicle capable of automatic driving (including automatic driving support), it goes without saying that the vehicle travels on the road to reach the target position while satisfying the adaptability and responsiveness to the recognized latest driving environment. The comfort of the occupant until reaching the target position, the riding comfort, for example, the smoothness of the behavior change of the vehicle similar to that of an exemplary driver is emphasized.

However, in JP5306934B, due to the fact that the robot is a legged robot and the like, the operation of the legged robot is momentarily controlled in a form that preferentially reflects the evaluation result by the action purpose generation module having a short calculation cycle, and the target position is controlled. However, there is room for improvement in improving the smoothness of the behavior change of the controlled object (the comfort of the occupant's riding comfort).

The present invention has been made in consideration of the above problems, and can be appropriately applied to a trajectory generation process of an autonomous driving vehicle (including an automatic driving assistance vehicle), and obtains a stable trajectory output. It is an object of the present invention to provide a vehicle control device that enables it.

A vehicle control device according to the present invention is a vehicle control device for controlling a vehicle capable of autonomous driving, and at least a lane shape of a lane in which a vehicle is traveling is determined from the external world information detected by an external world sensor that detects external world information. An external world recognition unit that recognizes and generates external world recognition information including the recognized lane shape information, a first trajectory generation unit that generates a first trajectory in a first calculation cycle by using the external world recognition information, and the external world A second trajectory generation unit that generates a second trajectory longer than the first trajectory in a second computation cycle longer than the first computation cycle, using the recognition information, the first trajectory generation unit, and the second trajectory generation And a general control unit that controls the unit. Here, the external world recognition unit performs the recognition processing in a cycle that is equal to or shorter than the first calculation cycle, and the overall control unit determines when the external world recognition information used by the second trajectory generation unit is present. Causes the first trajectory generation unit to generate the first trajectory by referring to the external world recognition information and the second trajectory.

As described above, according to the present invention, when the first trajectory that is the lower layer uses the latest external world recognition information and the second trajectory that is generated by the second trajectory generating section of the upper layer to perform trajectory optimization, Even if the external world information (environmental information) has not changed, the external world information that is different from the external world information used by the second trajectory generation unit, which is the upper layer, is used due to a recognition error of the external world recognition unit. In that case, the second trajectory is not the optimum solution for the latest external world information, and an unnecessary multimodal property is generated in the evaluation function for evaluating the first trajectory, and the output trajectory (first trajectory) is May become unstable.

In reality, the lane shape formed of white lines, stop lines, and the like carries a recognition error (recognition noise), but is inherently static (does not change or does not move), and is the upper layer of the second trajectory generation unit. By using the external world recognition information that is partly (lane shape information) same as the external world recognition information used by, the lower layer first trajectory generation unit refers to the lower layer first trajectory generation unit and refers to the upper layer trajectory It is possible to suppress the mismatch of the external information (environmental information) with the second orbit, which is a stable trajectory output.

In this case, the outside world recognition information includes, in addition to the static outside world recognition information including the lane shape information, a dynamic outside world recognition information including traveling suppression source information that suppresses traveling of a vehicle including a traffic participant. When the first trajectory generation unit generates a trajectory, the integrated control unit dynamically updates the external environment recognition information with the latest external world recognition information. It is used by the trajectory generator.

According to this invention, when the first trajectory generation unit generates the first trajectory shorter than the second trajectory in the first computation cycle shorter than the second computation cycle of the second trajectory generation unit, it includes the traffic participants. Since the dynamic external world recognition information is generated by using the latest external world recognition information recognized in the first operation cycle or less, it is necessary to generate the first trajectory in which the responsiveness of the vehicle to traffic participants is not deteriorated. You can

In this case, the outside world recognition information includes, in addition to the lane shape information, sign/marking information that is static outside world recognition information that restricts the traveling of the own vehicle, and the overall control unit The sign/marking information may be used by the second trajectory generating unit, and the sign/marking information used by the second trajectory generating unit may be used by the first trajectory generating unit.

For example, the sign/marking information indicating the maximum speed, the sign/marking information for restricting the travel of the stop line, etc., which is used by the second track generation unit, is used by the first track generation unit, and thus is generated by the second track generation unit. The second trajectory that smoothly transitions or stops the vehicle speed can be used by the first trajectory generator.

In addition, a third trajectory generation unit that generates a third trajectory that is longer than the second trajectory by using the external environment recognition information in a third computation cycle that is longer than the second computation cycle and that is controlled by the central control unit. Further, the integrated control unit causes the first trajectory generation unit, the second trajectory generation unit, and the third trajectory generation unit to simultaneously start generation of each trajectory when starting the automatic operation. Before the second trajectory is generated, the first trajectory generation unit is caused to generate the first trajectory by using the latest external environment recognition information, and when the second trajectory is generated, the The first trajectory is generated by causing the first trajectory to be generated by using the external environment recognition information including the lane shape used by the second trajectory and the second trajectory generator, and the third trajectory is generated. In this case, the first trajectory may be generated by causing the first trajectory generation unit to use the external environment recognition information including the third trajectory and the lane shape information used by the third trajectory generation unit. Good.

According to the present invention, when the automatic operation is turned on, the first trajectory generation unit, the second trajectory generation unit, and the third trajectory generation unit first generate the first trajectory generation unit having the shortest calculation cycle. The vehicle is controlled on the first track, and then the vehicle is controlled on the first track in consideration of the second track generated by the second track generator having a medium calculation cycle, and then the calculation cycle is the longest. Generating the second trajectory in consideration of the third trajectory generated by the third trajectory generator, generating the first trajectory in consideration of the second trajectory generated in consideration of the third trajectory, and generating The vehicle is controlled on the above-described first track.

Therefore, when the automatic driving is turned on, the automatic driving can be immediately started, and the automatic driving can be gradually and gradually changed into the automatic driving in consideration of ride comfort and comfort. In this case, it is possible to suppress a mismatch between the upper layer orbit referenced by the lower layer orbit generation unit and the external world information (environmental information), and perform stable orbit output.

Here, in the outside world recognition information, in addition to the lane shape information, or the lane shape information and the sign/marking information, traveling suppression source information that suppresses traveling of the own vehicle is included, and The integrated control unit may cause the first and second trajectory generation units to use the traveling suppression source information without using the third trajectory generation unit.

According to the present invention, the riding comfort (smoothness of changes in vehicle behavior) is emphasized on the third track, which is a relatively long-term track, and is recognized on the second track, which is a medium-term track, and the first track, which is a short-term track. Since the adaptability and responsiveness to the external environment are emphasized, the adaptability to the external environment can be improved by using the travel suppression source information that suppresses the travel of the own vehicle when the first track and the second track are generated. While ensuring responsiveness, the vehicle is controlled by the first track that indirectly uses the third track generated using the lane shape, thereby performing automatic driving in consideration of ride comfort (comfort). be able to.

Furthermore, the lane shape information may be information recognized from a lane prescription provided on a road surface, and the travel restraint source information may be information including an obstacle, a traffic participant, or a traffic light color. ..

According to the present invention, the third track generation unit can generate a track with emphasis on ride comfort based on the information recognized from the lane prescription provided on the road surface, and the first track generation unit and the second track generation unit It is possible to generate a trajectory with emphasis on adaptability/responsiveness based on information including obstacles, traffic participants, or traffic light colors.

Furthermore, the lane defining object may include a lane mark or a lane departure prevention member.

According to this invention, the outside world recognition unit can recognize the lane shape of the lane in which the vehicle travels from the lane mark or a lane deviation prevention member such as a curb or guardrail.

It is a schematic block diagram of a vehicle equipped with a vehicle control device according to this embodiment. It is the block diagram which extracted the structure of the principal part in FIG. It is an illustration figure of an environment map. It is a flow chart provided for explanation of operation of the vehicle control device (manual operation mode). It is a flow chart (1st example) with which operation explanation (automatic operation mode) of a vehicle control device is offered. 4 is a time chart (first embodiment) used for explaining the operation of the vehicle control device. It is a flowchart (2nd Example) with which an operation|movement description (automatic driving mode) of a vehicle control apparatus is offered. 8A is a time chart of a comparative example, and FIG. 8B is a time chart according to the embodiment (first example) in which the time chart of FIG. 6 is deformed and drawn. 9A is a time chart of the same comparative example as FIG. 8A, and FIG. 9B is a time chart according to an embodiment (second example) in which the time chart of FIG. 6 is deformed and drawn.

Hereinafter, a vehicle control device according to the present invention will be described in relation to a vehicle in which the vehicle control device is mounted, with reference to the accompanying drawings, by way of preferred embodiments.

[Configuration of vehicle 10]
FIG. 1 is a block diagram showing a schematic configuration of a vehicle (also referred to as an own vehicle or an own vehicle) 10 equipped with a vehicle control device 12 according to this embodiment.

The vehicle 10 includes a vehicle control device 12 and, in addition to the vehicle control device 12, includes an input device and an output device that are connected to the vehicle control device 12 via communication lines.

As the input device, an external sensor 14, a navigation device 16, a vehicle sensor 18, a communication device 20, an automatic operation switch (automatic operation SW) 22, and an operation detection sensor 26 connected to an operation device 24. Prepare

The output device includes an actuator 27 having a driving force device 28 that drives wheels (not shown), a steering device 30 that steers the wheels, and a braking device 32 that brakes the wheels.

The navigation device 16 and the communication device 20 can also be used as input/output devices (human interface, transceiver).

[Configuration of input/output device connected to vehicle control device 12]
The external environment sensor 14 includes a plurality of cameras 33 and a plurality of radars 34 that acquire information on the external environment of the vehicle 10 (around 360 degrees around the front, the rear, the side, etc.), and the acquired external environment information on the vehicle 10 is a vehicle control device. Output to 12. The external sensor 14 may further include a plurality of LIDARs (light detection and distance measurement).

The navigation device 16 detects and specifies the current position of the vehicle 10 by using a satellite positioning device or the like, and has a touch panel display, a speaker, and a microphone as a user interface, and specifies from the current position or a position specified by the user. The route to the destination is calculated and output to the vehicle control device 12. The route calculated by the navigation device 16 is stored in the route information storage unit 44 of the storage device 40 as route information.

The vehicle sensor 18 is a speed (vehicle speed) sensor that detects the speed (vehicle speed) of the vehicle 10, an acceleration sensor that detects acceleration, a lateral G sensor that detects lateral G, and a yaw rate sensor that detects an angular velocity around the vertical axis of the vehicle 10. Outputs each detection signal to the vehicle control device 12 including an azimuth sensor that detects the direction of the vehicle 10 and a gradient sensor that detects the gradient of the vehicle 10. These detection signals are stored in the vehicle state information storage unit 46 of the storage device 40 as the vehicle state information Ivh for each calculation cycle Toc described later.

The communication device 20 communicates with a roadside device, another vehicle, a server, etc., and receives or transmits information about a traffic light, information about another vehicle, probe information, updated map information, and the like. The map information is stored in the navigation device 16 and also in the map information storage unit 42 of the storage device 40 as map information.

The operation device 24 includes an accelerator pedal, a steering wheel (handle), a brake pedal, a shift lever, a direction indicating lever, and the like. The operation device 24 is provided with an operation detection sensor 26 for detecting the presence/absence of operation by the driver, the operation amount, and the operation position.

The operation detection sensor 26 outputs to the vehicle control unit 110, as detection results, an accelerator depression (opening) amount, a steering wheel operation (steering) amount, a brake depression amount, a shift position, a left/right turning direction, and the like.

The automatic operation switch (automatic operation switch) 22 is provided, for example, on an instrument panel, and is manually operated by a user such as a driver to switch between a non-automatic operation mode (manual operation mode) and an automatic operation mode. It is a push button switch.

The automatic driving mode is a driving mode in which the vehicle 10 travels under the control of the vehicle control device 12 while the driver does not operate the operation device 24 such as the accelerator pedal, the steering wheel, or the brake pedal. Is a driving mode in which a part or all of the driving force device 28, the steering device 30, and the braking device 32 are controlled based on an action plan (a short-term trajectory St, a medium-term trajectory Mt, and a long-term trajectory Lt described later).

When the driver starts operating the operation device 24 such as the accelerator pedal, the steering wheel, or the brake pedal during the automatic driving mode, the automatic driving mode is automatically canceled and the non-automatic driving mode (manual driving Mode).

Here, even in the manual operation mode, it is possible to implement certain driving support functions such as a known ACC (Adaptive Cruise Control) function and a LKAS (Lane Keep Assist System) function.

Further, the automatic operation switch 22 described above may be of a touch type or a voice input type.

The driving force device 28 includes a driving force ECU and a driving source of the vehicle 10 such as an engine and/or a driving motor. The driving force device 28 generates a traveling driving force (torque) for the vehicle 10 to travel according to the vehicle control value Cvh input from the vehicle control unit 110, and transmits it to the wheels via the transmission or directly.

The steering device 30 includes an EPS (electric power steering system) ECU and an EPS device. The steering device 30 changes the direction of the wheels (steering wheels) according to the vehicle control value Cvh input from the vehicle control unit 110.

The braking device 32 is, for example, an electric servo brake that also uses a hydraulic brake, and includes a brake ECU and a brake actuator.

The braking device 32 brakes the wheels according to the vehicle control value Cvh input from the vehicle control unit 110.

The steering of the vehicle 10 can be performed by changing the torque distribution or the braking force distribution to the left and right wheels.

[Configuration of vehicle control device 12]
The vehicle control device 12 includes one or a plurality of ECUs (electronic control units), and includes a storage device 40 and the like in addition to various function realizing units. In this embodiment, the function realizing unit is a software function unit whose function is realized by the CPU (central processing unit) executing the program stored in the storage device 40. It can also be realized by a hardware function unit.

FIG. 2 is a block diagram showing a configuration of a main part of the vehicle control device 12 according to this embodiment, which is extracted from FIG.

The vehicle control device 12 includes an external environment recognition unit 51, a recognition result reception unit 52, and a local environment map generation unit, in addition to the storage device 40 (FIG. 1) and the vehicle control unit 110 as a function realization unit (function realization module). (Also referred to as an environment map generation unit) 54, a long-term orbit generation unit 71, a medium-term orbit generation unit 72, a short-term orbit generation unit 73, and an overall control unit (task) that controls task synchronization. (Synchronization module) 70.

In the vehicle control device 12, the external world recognition unit 51 includes an external world recognition information Iprs which is static (changes or does not move) and dynamic (changes or may move) external world recognition information Iprd. Generate Ipr at the same time.

When generating the static external world recognition information Iprs, the external world recognition unit 51 refers to the own vehicle state information Ivh from the vehicle control unit 110, and further, external world information (image information from the camera 33, etc. in the external world sensor 14). Based on the above), the lane marks (white lines, etc.) on both sides (right and left sides) of the vehicle 10 at that position are recognized, and the distance to the stop line such as an intersection (the number of meters to the stop line). ), a drivable area (a plane area excluding guardrails and curbs without worrying about the lane mark), and the like are generated as static external environment recognition information Iprs and transmitted (output) to the recognition result receiving unit 52. ..

The static external environment recognition information Iprs includes lane shape information Iprsl indicating the shape of a lane (straight lane, curved lane, slope lane, etc.) defined by lane regulations such as lane marks, curbs, and guardrails. Includes signs/marking information Iprsm including signs (maximum speed sign prohibiting overtaking signs, etc.) and signs (stop line display, course change prohibiting signs, etc.) that restrict the traveling of the vehicle 10 included in the outside world information (image information) Be done. The curbstone and the guardrail function as a lane departure prevention member.

When generating the dynamic outside world recognition information Iprd, the outside world recognition unit 51 refers to the own vehicle state information Ivh and further based on the outside world information from the camera 33 or the like, obstacles (falling objects, parked vehicles, parked vehicles, Recognize the travel restraint source information iprdc such as traffic color (blue (green), yellow (orange), red}) of traffic participants (including animals, people, other vehicles including motorcycles), and traffic lights It is generated as the external world recognition information Iprd and transmitted (output) to the recognition result receiving unit 52.

The external world recognition unit 51 recognizes the external world recognition information Ipr (Ipr=Iprs+Iprd) in a time shorter than the calculation cycle Toc, and transmits (outputs) the recognition result reception unit 52.

In this case, the recognition result receiving unit 52 responds to the operation command Aa from the overall control unit 70 with the outside world recognition information Ipr (Ipr=Iprs+Iprd) received from the outside world recognition unit 51 within the operation cycle Toc. Output to the integrated control unit 70.

The overall control unit 70 stores the external world recognition information Ipr (Ipr=Iprs+Iprd) in the storage device 40.

Here, the calculation cycle (also referred to as a reference cycle or a reference calculation cycle) Toc is a reference calculation cycle in the vehicle control device 12, and is set to a value of, for example, several tens of ms.

The local environment map generation unit 54 refers to (aggregates) the own vehicle state information Ivh and the external environment recognition information Ipr in response to the operation command Ab from the overall control unit 70, and sets the environment map information (local) in the operation cycle Toc. It is also referred to as environment map information.) Generated as Iem and output to the integrated control unit 70.

The environment map information Iem is generally information in which the own vehicle state information Ivh is combined with the outside world recognition information Ipr. The environment map information Iem is stored in the environment map information storage unit (also referred to as local environment map information storage unit) 47 of the storage device 40.

FIG. 3 illustrates an example environment map (also referred to as a local environment map) Lmap stored as the environment map information Iem.

Here, the vehicle state information Ivh is information obtained from the vehicle control unit 110, and basically, the lane L (right lane mark Lmr) at the reference point Bp of the vehicle 10, for example, the middle point of the rear wheel axle. And the left-side lane mark Lml.) from the center line (virtual line) CL of the offset amount (position) OS, and the posture angle (angle) between the center line CL and the nose direction nd of the vehicle 10 ( Also referred to as azimuth angle.) θz, speed vs, acceleration va, curvature ρ of traveling line, yaw rate γ, steering angle δst, and the like. The offset amount OS may be coordinates {x (the direction of the traveling road and the longitudinal direction) y (the direction orthogonal to the traveling road and the lateral direction)} from the reference position (arbitrary).

That is, the own vehicle state information Ivh is the latest information at that point in time of a track point sequence Pj {see Formula (2)} described later, as shown in Formula (1) below.
Ivh=Ivh(x, y, θz, vs, va, ρ, γ, δst) (1)
Pj
=Pj(x, y, θz, vs, va, ρ, γ, δst) t=1, 2,... T
…(2)

The trajectory point sequence Pj is corrected until the trajectory point sequence candidate Pcj (x, y, θz, vs, va, ρ, γ, δst) t=1, 2,... T is positively evaluated. Then, the trajectory point sequence Pj (x, y, θz, vs, va, ρ, γ, δst) t=1, 2,... T is set as the output trajectory. t corresponds to a time that is an integer fraction of the calculation cycle Toc (may be changed according to the velocity vs.), 1 is a first point, and T is, for example, a point at the first sec. Corresponding to the time length of the orbit.

In FIG. 3, the lane L (the right lane mark Lmr and the left lane mark Lml) is recognized by the external world recognition unit 51 from the image information from the camera 33 (known lane mark detection, bird's-eye view conversion, and curve approximation processing). This is the recognition information Ipr.

In this way, the environment map information Iem (environment map Lmap) is generated by merging the own vehicle state information Ivh and the external environment recognition information Ipr, and is based on the own vehicle position in the direction in which the own vehicle 10 is traveling. This is information indicating the surroundings of the road (lane mark Lm) and the like (the surroundings of the own vehicle).

In the local environment map generation unit 54, for example, the lane center line CL is generated as an optimum traveling line in the case of a straight lane, and if it is a curved lane, so-called out-in-out with respect to the lane center line CL. The travel line is generated as the optimum travel line. This optimum travel line is included in the environment map information Iem (environment map Lmap).

The lane L (the shape of the lane L) is defined by the lane markings such as the curb Es and the guardrail Gr in addition to the lane mark Lm.

Returning to FIG. 2, the long-term trajectory generation unit 71 responds to the operation command Ac from the overall control unit 70, and the environment map information Iem including the static external world recognition information Iprs excluding the dynamic external world recognition information Iprd, With reference to the vehicle state information Ivh and the road map (curvature of the curve, etc.) stored in the map information storage unit 42, for example, a long-term trajectory Lt is generated at a calculation cycle of 9×Toc, and the overall control unit 70 is generated. Output. The long-term trajectory Lt is stored in the trajectory information storage unit 48 of the storage device 40 as trajectory information It.

That is, the long-term trajectory generation unit 71 is a trajectory for performing vehicle control that emphasizes the riding comfort/comfort of the vehicle 10 (does not perform sudden steering or sudden acceleration/deceleration), for example, a model driver familiar with driving drives the vehicle. Is a trajectory corresponding to the trajectory, and the static external environment recognition information Iprs is used without using the dynamic external world recognition information Iprd, and the operation cycle is a relatively long cycle, for example, a long cycle Tl of about several hundred ms. At (Tl=9×Toc), a long-term trajectory (also referred to as a 10-sec trajectory) Lt corresponding to a relatively long time (long distance), for example, a traveling time of about 10 seconds is generated.

In response to the operation command Ad from the overall control unit 70, the medium-term trajectory generation unit 72 includes environment map information Iem (including dynamic external world recognition information Iprd and static external world recognition information Iprs ) and own vehicle state information. With reference to Ivh and the long-term orbit Lt, the medium-term orbit Mt is generated at a calculation cycle of 3×Toc, and the generated medium-term orbit Mt is output to the overall control unit 70. The medium-term trajectory Mt is stored in the trajectory information storage unit 48 as trajectory information It .

The medium-term trajectory generation unit 72 bypasses the parked vehicle or the like when the external world recognition unit 51 finds an obstacle such as a parked vehicle (included in the dynamic external world recognition information Iprd) in front of the lane L, for example. This is a trajectory to be made (a trajectory including lane change if necessary when there are a plurality of lanes on one side), and a calculation cycle is a cycle relatively shorter than the long cycle Tl, for example, a medium cycle Tm of about 100 to 10 ms ( At Tm=3×Toc, a medium-term orbit (also referred to as 5 sec orbit) Mt corresponding to a relatively short time (short distance), for example, a traveling time of several seconds is generated.

When the dynamic map recognition information Iprd is not included in the environment map information Iem when the medium-term trajectory Mt is generated, as a result, the medium-term trajectory Mt substantially matches the long-term trajectory Lt.

The short-term trajectory generation unit 73 responds to the calculation command Ae from the overall control unit 70, and includes the environment map information Iem (including the dynamic external world recognition information Iprd and the static external world recognition information Iprs) and the vehicle state information. Ivh and the medium-term trajectory Mt generated by referring to the long-term trajectory Lt are referred to, and the short-term trajectory St corresponding to the vehicle dynamics of the host vehicle 10 is generated and generated at the shortest calculation cycle Toc among the three trajectory generators. The short-term trajectory St is output to the overall control unit 70 and simultaneously to the vehicle control unit 110.

The vehicle control unit 110 controls the actuator 27 based on the short-term trajectory St. The short-term trajectory St is stored in the trajectory information storage unit 48 as trajectory information It.

When the dynamic external world recognition information Iprd is not included in the environment map information Iem when the short-term trajectory St is generated, as a result, the short-term trajectory St is generated by referring to the long-term trajectory Lt. It roughly matches Mt.

In this way, the short-term trajectory generation unit 73 uses a calculation cycle that is relatively shorter than the long cycle Tl and the middle cycle Tm, for example, a short cycle Ts (Ts=Toc) of about several tens of ms, and the relative distance to travel from now on. A short-term track (hereinafter referred to as a 1-sec track) St corresponding to a running time of a short time (short distance), for example, about 1 second is generated.

As the short-term track St, a vertical position x, a horizontal position y, a posture angle θz, a velocity vs, an acceleration va, and a steering angle δst (of the vehicle 10), which are substantially along the center line CL of the lane mark, are provided for each short cycle Ts. The steering angle δ can be calculated by taking the gear ratio into consideration in the steering angle δst of the steering wheel.) and the like, based on the vehicle command value, the track point sequence Pj(x, y, θz, vs, va, δst. ) {Refer to the above formula (2). } Is generated.

In practice, before the final trajectory point sequence Pj is generated, the short-term trajectory generation unit 73 generates a plurality of trajectory point sequence candidates Pcj (computation cycle: Toc/5) for each short cycle Ts (Ts=Toc). Is generated. The generated trajectory point sequence candidate Pcj is necessary according to the evaluation result after the trajectory is evaluated by the short-term trajectory generator 73 based on the vehicle dynamics and the like within the same short cycle Ts, as described later. If corrected, the trajectory point sequence Pj as an output trajectory corresponding to the short-term trajectory St is generated.

The vehicle control unit 110 sets the track control point Pj to the vehicle control value so that the vehicle 10 travels along the input short-term track St, that is, the track point sequence Pj generated and input at the calculation cycle Toc/5. It is converted into Cvh and output to the driving force device 28, the steering device 30, and the braking device 32.

[Description of Operation of Embodiment]
[Explanation by flow chart]
The operation of the vehicle control device 12 basically configured as described above will be described in detail with reference to the flowchart of the manual operation mode of FIG. 4 and the flowchart of the automatic operation mode of FIG. The execution subject of the program according to the flowchart is the overall control unit 70 or the like of the vehicle control device 12.

(A) Description of Flowchart of Manual Operation Mode in FIG. 4 In step S1, the overall control unit 70 sends the calculation result Aa requesting the recognition result receiving unit 52 to receive the external world recognition information Ipr, and An operation command Ab requesting the generation of the environment map information Iem is sent to the local environment map generator 54.

In this case, the outside world recognizing unit 51 receives outside world information (image information, radar emission/reflection information, and light irradiation/reflection) from the camera 33, the radar 34, and LIDAR (not shown) in the external sensor 14 within the calculation cycle Toc. The external environment recognition information Ipr including static external environment recognition information Iprs (lane shape information Iprsl, sign/marking information Iprsm) and dynamic external world recognition information Iprd (driving suppression source information Iprdc), which is recognized based on the information) is generated. Then, it is transmitted (output) to the recognition result receiving unit 52.

The recognition result receiving unit 52 outputs the external world recognition information Ipr (Iprs+Iprd) to the overall control unit 70 in response to the operation command Aa.

On the other hand, the local environment map generation unit 54 merges the own vehicle state information Ivh with the external environment recognition information Ipr within the calculation cycle Toc in synchronization with the calculation command Ab to generate the environment map Lmap shown in FIG. The generated environment map information Iem is generated and sent to the overall control unit 70.

In step S<b>2, the overall control unit 70 acquires the outside world recognition information Ipr and the environment map information Iem and stores them in the storage device 40.

Note that the external world recognition information Ipr is not generated by the external world recognition unit 51 in the first calculation command Ab in step S1, so the environment map information Iem is generated after the second generation request in step S1. That is, the environment map information Iem is generated based on the external environment recognition information Ipr generated in the previous calculation cycle Toc and the latest own vehicle state information Ivh.

Next, in step S3, the overall control unit 70 determines whether or not the automatic driving switch 22 is set to the on-state automatic driving mode.

When the automatic operation switch 22 is set to the off-state non-automatic operation mode (step S3: NO), the generation processing of the external world recognition information Ipr and the environment map information Iem in steps S1 and S2 is repeated.

When the automatic operation switch 22 is set to the on-state automatic operation mode (step S3: YES), the automatic operation mode is instantaneously switched in step S4 {transition from non-automatic operation mode to automatic operation mode (shift) It is also said to be done. }.

[Explanation by Flowchart (First Embodiment)]
(B) Description of the operation of the first embodiment of the vehicle control device 12 with reference to the flowchart relating to the automatic driving mode (first embodiment) of FIG. Calculation command Aa and environment map information requesting the reception unit 52, the local environment map generation unit 54, the short-term trajectory generation unit 73, the medium-term trajectory generation unit 72, and the long-term trajectory generation unit 71 to receive the external environment recognition information Ipr. An operation command Ab requesting generation of Iem, an operation command Ac requesting generation of a short-term trajectory St, an operation command Ad requesting generation of a medium-term trajectory Mt, and an operation command Ae requesting generation of a long-term trajectory Lt are simultaneously transmitted. ..

In this case, the short-term trajectory generation unit 73 generates the short-term trajectory St based on the environment map information Iem generated immediately before and the own vehicle state information Ivh obtained at the start time in synchronization with the first calculation cycle Toc. can do.

The medium-term trajectory generation unit 72 can generate the first medium-term trajectory Mt after the computation cycle 3×Toc, which is the triple computation cycle, including the first computation cycle Toc, and thereafter every computation cycle 3×Toc. A mid-term trajectory Mt is generated.

The long-term trajectory generation unit 71 can generate the first long-term trajectory Lt after the computation cycle 9×Toc, which is a nine-fold computation cycle, including the first computation cycle Toc, and thereafter every computation cycle 9×Toc. Generate a long-term trajectory Lt.

Therefore, after the next step S12, the short-term trajectory generation process in the short-term trajectory generator 73 according to the main part of the present invention will be described.

In this embodiment, basically, in the external world recognition information Ipr used when the short-term trajectory generation unit 73 generates the short-term trajectory St, the static external world recognition information Iprs and the medium-term trajectory generation unit 72 identify the medium-term trajectory Mt. The purpose is to avoid the mismatch (mismatch) of the static external world recognition information Iprs used when generating, and to generate the stable short-term trajectory St. Note that when the short-term trajectory generation unit 73 generates the short-term trajectory St, the latest dynamic external environment recognition information Iprd is used.

Originally, the static external world recognition information Iprs such as the lane mark and the stop line does not change within the calculation cycle (less than Toc) by the external world recognition unit 51. However, the lane mark or the stop line in the image information may be momentarily interrupted due to the presence of backlight, a puddle, or the like.

In such a case, the latest static external world recognition information Iprs used by the short-term trajectory generation unit 73 and the static external world recognition information Iprs used by the medium-term trajectory generation unit 72 do not match.

Therefore, when the short-term trajectory generation unit 73 generates the short-term trajectory St, whether or not there is the static external world recognition information Iprs used when the medium-term trajectory generation unit 72 generates the medium-term trajectory Mt in step S12. Determine whether.

If not (step S12: NO), in step S13, the short-term trajectory generation unit 73 uses the latest static external environment recognition information Iprs, dynamic external world recognition information Iprd, and own vehicle state information Ivh. A short-term trajectory St is generated and automatic driving is executed.

Next, in step S14, it is determined whether or not the automatic operation switch 22 is set to the off-state non-automatic operation mode (manual operation mode).

When the automatic operation switch 22 is set to the automatic operation mode in the ON state (step S14: NO), the processes of step S11 and subsequent steps are repeated, and normally, the determination of step S12 is made after the elapse of the calculation cycle of 3×Toc. If the determination is affirmative (step S12: YES), the process of generating the short-term trajectory St in step S15 is executed.

In the generation process of the short-term trajectory St in step S15, the short-term trajectory St is utilized by using the static external environment recognition information Iprs, the latest dynamic external environment recognition information Iprd, and the own vehicle state information Ivh used by the medium-term trajectory generation unit 72. Is generated and the automatic operation is continued.

In the determination of step S14, if the automatic operation switch 22 is set to the manual operation mode in the off state (step S14: YES), the manual operation mode is switched to in step S16.

[Explanation by time chart]
Next, the operation of the vehicle control device 12 according to the first embodiment will be described with reference to the time chart of FIG. 6 (first embodiment).

In FIG. 6, at time t0, the driver operates the automatic operation switch 22 to switch from the manual operation mode (automatic operation off state) to the automatic operation mode (automatic operation on state).

At time t-2 (the time at the leftmost end in FIG. 6) before time t0, the overall control unit 70, near the start of the process Pro1 corresponding to the calculation cycle Toc, the recognition result receiving unit 52 (FIG. 1). 2), a calculation command Aa (not shown in FIG. 6) for requesting reception of the external environment recognition information Ipr is sent, and a calculation command Ab for requesting the local environment map generation unit 54 to generate the environment map information Iem. It is sent (corresponding to step S1).

The local environment map generation unit 54 generates the environment map information Iem1 at the time of the substantially calculation cycle Toc in response to the calculation command Ab at the time point t-2 and outputs it to the overall control unit 70.

Since the environment map information Iem1 is the combined information of the outside world recognition information Ipr and the vehicle state information Ivh, the outside world recognition unit responds to the operation command Aa before the processing Pro1 for generating the environment map information Iem ends. Although the external environment recognition information Ipr received from 51 is sent from the recognition result receiving section 52 to the local environment map generating section 54 via the overall control section 70, the calculation command Aa and the external environment recognition are performed due to space limitations. The recognition process of the information Ipr is omitted.

As described above, it is assumed that the automatic operation switch 22 is operated in the ON state at the time point t0 before the time point t1 at which the process Pro2 of the overall control unit 70 ends (step S3: YES).

The vehicle control unit 110 sends the host vehicle state information Ivh1 to the overall control unit 70 in response to the automatic driving switch 22 being turned on at the time point t1 of the end point of the process Pro2 of the overall control unit 70. ..

In this case, at the end time point t3 of the process Pro3 of the calculation cycle Toc, the vehicle control value generated by the vehicle control unit 110 only at the short-term track St1 input from the short-term track generation unit 73 at the time point t2 immediately before the time point t3. The actuator 27 (driving force device 28, steering device 30, and braking device 32) is controlled based on Cvh, and automatic operation is performed (corresponding to the first step S13).

On the other hand, in response to the automatic operation switch 22 being operated to the ON state, the central control unit 70 near the start point of the process Pro3, the short-term trajectory generation unit 73, the medium-term trajectory generation unit 72, and the short-term trajectory generation unit 72. Calculation commands Ae, Ad, and Ac requesting generation of the short-term orbit St3, the medium-term orbit Mt1, and the long-term orbit Lt1 are sent to the long-term orbit generator 71 (corresponding to step S11).

Then, the medium-term trajectory Mt1 is generated by the processing Pro1 (middle-term trajectory generation unit 72) during the processing Pro5 (overall control unit 70) when approximately three times the calculation cycle Toc has elapsed, and the short-term trajectory is generated via the overall control unit 70. It is sent to the generation unit 73.

In this case, in the process Pro6 (overall control unit 70) in the next operation cycle Toc, in other words, in the process Pro4 of the short-term trajectory generation unit 73, the medium-term trajectory Mt1 of the 5 sec trajectory is taken into consideration. The short-term track St4 that makes the traveling line of the vehicle 10 smoother is generated and output to the overall control unit 70 and the vehicle control unit 110.

In addition, considering the medium-term trajectory Mt1 means that when the short-term trajectory generator 73 generates the short-term trajectory St4, the surrounding environment based on the current position velocity vs of the vehicle 10, the acceleration va, the yaw rate γ, and the steering angle δst. In consideration of the above, a plurality of track point sequence candidates Pcj for selecting the short-term track St4 from the current position (start point) of the vehicle 10 to the target point (end point) after 1 [sec] are generated.

Then, each trajectory point sequence candidate Pcj of the generated short-term trajectory St4 is evaluated based on an evaluation function described later with respect to the trajectory point sequence Pj of the medium-term trajectory Mt1, and an environment map Lmap (this environment map Lmap is a medium-term trajectory The static external environment recognition information Iprs2 in the environment map information Iem2 acquired by the generation unit 72 from the local environment map generation unit 54 via the recognition result reception unit 52 at time t1, and the short-term trajectory generation unit 73 performs processing Pro4. Note that this corresponds to the dynamic external world recognition information Iprd5 in the latest environment map information Iem5 acquired from the local environment map generation unit 54 via the recognition result reception unit 52 at the start point). Meaning that a trajectory point sequence candidate Pcj that has been evaluated based on another evaluation function described later and has a high evaluation is selected, and a short-term trajectory St4 in which the selected trajectory point sequence candidate Pcj is the trajectory point sequence Pj is generated. Is.

The track point sequence Pj of the short-term track St4 is converted into the vehicle control value Cvh at the cycle of the calculation cycle Toc/5 by the vehicle control unit 110, and is output to the actuator 27, so that the adaptability to the external environment and the responsiveness of the vehicle 10 are improved. The automatic driving that is considered is performed (in FIG. 6, a period from time t4 to time t5, which is added as "Automatic driving at St+Mt").

In this case, the evaluation function with the orbital point sequence Pj of the medium-term orbit Mt1 is that each element (position x, y, at the corresponding point of the orbital point sequence Pj of the medium-term orbit Mt1 and each orbital point sequence candidate Pcj of the short-term orbit St4. The smaller the deviation of the velocity vs, the steering angle δst, etc. (deviation related to the vehicle control value Cvh), the higher the evaluation is set, and the environment map Lmap (static external environment recognition information Iprs2 and dynamic external world recognition information) is set. An optimal travel line of the lane L or the vehicle 10 generated from Iprd5 and the vehicle state information Ivh4, for example, a lane center line CL for a straight lane and an out-in-out travel line for a curved lane. The evaluation function for) is higher the smaller the deviation (positional deviation related to the lane L) between the positions x and y of each track point sequence candidate Pcj of the short-term trajectory St4 and the optimum travel line of the environment map Lmap, and the like, the higher the evaluation. Is set to be. The trajectory point sequence candidate Pcj having the highest weighted sum of the evaluation values of both evaluation functions is set to the trajectory point sequence Pj of the short-term trajectory St4.

Next, a long-term trajectory Lt1 is generated by the process Pro1 of the long-term trajectory generation unit 71 during the process Pro11 by the overall control unit 70 when approximately nine times the computation period Toc has elapsed from the time point t1, and the process Pro15 (control time during the computation period Toc) In the control unit 70), the short-term trajectory St13 of the vehicle 10 in consideration of the long-term trajectory Lt1 of the 10-sec trajectory and the medium-term trajectory Mt3 of the 5-sec trajectory is generated by the process Pro13 of the short-term trajectory generator 73.

The short-term orbit St13 considers the long-term orbit Lt1 and the medium-term orbit Mt3 means that, for example, when the medium-term orbit generator 72 generates the medium-term orbit Mt4 in the process Pro4, it is the same as that described in the generation process of the short-term orbit St4. , A candidate for a plurality of medium-term trajectories Mt4 composed of a plurality of trajectory point sequence candidates Pcj is generated, and each generated trajectory point sequence candidate Pcj is evaluated based on an evaluation function with the trajectory point sequence Pj of the long-term trajectory Lt1, and Evaluation based on the evaluation function of the environment map Lmap (in this case, it corresponds to the static external environment recognition information Iprs11 in the latest environment map information Iem11 and the dynamic external world recognition information Iprd11). Then, the trajectory point sequence candidate Pcj having a high total evaluation is set as the trajectory point sequence Pj of the medium-term trajectory Mt4.

Further, the medium-term orbit Mt4 generated by referring to the long-term orbit Lt1 and the plurality of candidates for the short-term orbit St13 are evaluated by the evaluation function as described above, and thus generated by referring to the long-term orbit Lt1. This means that the short-term orbit St13 is generated with reference to the mid-term orbit Mt4 that was generated.

After the time point t5, during a period noted as “Automatic driving with St+Mt+Lt”, automatic driving with extremely excellent adaptability and responsiveness is executed, which is close to the driving feeling of the driver and the ride comfort is fully taken into consideration.

[Explanation by Flowchart (Second Embodiment)]
(C) Description of the operation of the second embodiment of the vehicle control device 12 with reference to the flowchart relating to the automatic driving mode (second embodiment) of FIG. 7. The flowchart of FIG. 7 corresponds to the flowchart shown in FIG. In comparison, the only difference is that steps S12a, S12B, and S15a are added.

That is, when the short-term trajectory generation unit 73 generates the short-term trajectory St, in step S12a, whether or not the environment region to be automatically driven is the transient region is the static external environment recognition information Iprs and/or the map information storage unit 42. It is determined based on the road map (curvature curvature determination, etc.) stored in. Specifically, it is confirmed whether the change in the steering angle δst is equal to or more than a threshold value when entering a curve, turning left or right, and the like, and it is determined that the transient region is equal to or more than the threshold value (step S12a: YES). In step S12, the processes after step S12 described above are executed.

On the other hand, when it is determined that the change in the steering angle δst is less than the threshold value and is not in the transient region when the vehicle is traveling straight or when the vehicle is traveling in a curve with good visibility (step S12a: NO), the short-term trajectory generator 73 is used. At the time of generating the short-term trajectory St, at step S12b, it is determined whether or not there is the static external world recognition information Iprs used when the long-term trajectory generation unit 71 generates the long-term trajectory Lt.

If not (step S12b: NO), the above-described processing of step S12 and thereafter is performed.

If there is (step S12b: YES), in the process of generating the short-term orbit St in step S15a, the static external environment recognition information Iprs used by the long-term orbit generation unit 71, the latest dynamic external world recognition information Iprd, and the self-recognition information Iprd. The vehicle state information Ivh is used to generate the short-term track St, and the automatic driving is continued until the determination in step S14 becomes affirmative (step S14: YES).

In the second embodiment, since the short-term trajectory St is very similar to the long-term trajectory Lt when it is not in the transient region, it is possible to perform extremely stable trajectory output in which the mismatch of external environment information (environmental information) is suppressed.

Explaining an example with the time chart of FIG. 6, in the process Pro13 of the short-term orbit generation unit 73, the static external world recognition information Iprs2 acquired by the long-term orbit generation unit 71 at time t1 is used as the static external world recognition information Iprs. doing. By the way, in the second embodiment, also in the process Pro4 of the medium-term orbit generation unit 72, the static external world recognition information Iprs2 acquired by the long-term orbit generation unit 71 at time t1 is used as the static external world recognition information Iprs. There is.

[Summary]
8A is a time chart according to a comparative example, and FIG. 8B is a deformed and drawn first embodiment (used by the short-term trajectory generation unit 73 for convenience of understanding in the time chart shown in FIG. 6. The static external world recognition information Iprs is the same as the static external world recognition information Iprs used by the medium-term trajectory generation unit 72.). The time point t1 in FIG. 6 corresponds to the time point t11 in FIGS. 8A and 8B.

In the comparative example of FIG. 8A, the long-term trajectory generation unit 71, the medium-term trajectory generation unit 72, and the short-term trajectory generation unit 73 use the latest static external environment recognition information Iprs such as the lane shape information Iprsl. Generating an orbit.

For example, the short-term trajectory generation unit 73 generates the short-term trajectory St14 using the static external environment recognition information Iprs obtained at the time point t24 in the process Pro14.

The medium-term trajectory generation unit 72 generates a medium-term trajectory Mt5 (not shown) using the static external world recognition information Iprs and the dynamic external world recognition information Iprd obtained at the time point t23 in the process Pro5.

The long-term trajectory generation unit 71 uses the static external environment recognition information Iprs obtained at the time point t20 in the process Pro2 to generate the long-term trajectory Lt.

Thus, in the external environment recognition information Ipr used when the short-term trajectory generation unit 73 generates the short-term trajectory St, the long-term trajectory generation unit 71 and the medium-term trajectory generation unit 72 are long-term corresponding to the static external environment recognition information Iprs. There is a mismatch (mismatch) with the static external environment recognition information Iprs used when generating the trajectory Lt and the medium-term trajectory Mt.

In the first embodiment shown in FIG. 8B, the static external environment recognition information Iprs such as the lane shape information Iprsl is used in the lower layer short-term trajectory generation unit 73 that is used by the upper layer medium-term trajectory generation unit 72. I am trying to do it.

For example, in the process Pro14, the short-term trajectory generation unit 73 generates the short-term trajectory St14 using the static external environment recognition information Iprs acquired by the middle-term trajectory generation unit 72 at time t20.

The medium-term trajectory generation unit 72 generates a medium-term trajectory Mt5 (not shown) using the static external world recognition information Iprs and the dynamic external world recognition information Iprd obtained at the time point t23 in the process Pro5.

FIG. 9A is a time chart according to the same comparative example as shown in FIG. 8A, and FIG. 9B shows a time chart shown in FIG. 6 in a deformed and drawn second embodiment for convenience of understanding. It is such a time chart. The time point t1 in FIG. 6 corresponds to the time point t11 in FIGS. 9A and 9B.

In the second embodiment shown in FIG. 9B, it is premised that it is not in the transient region (step S12a: NO), and that the static external world recognition information Iprs used when generating the long-term trajectory Lt exists (step S12b: YES). , The external environment recognition information Iprs such as the lane shape information Iprsl used by the long-term orbit generation unit 71 as the upper layer is used by the medium-term orbit generation unit 72 and the short-term trajectory generation unit 73 as the lower layer. ing.

For example, in the process Pro14, the short-term trajectory generation unit 73 uses the static external environment recognition information Iprs acquired by the medium-term trajectory generation unit 72 at time t11 and the dynamic external world recognition information Iprd acquired by itself at time t24. Then, the short-term orbit St14 is generated.

Further, the medium-term trajectory generation unit 72 uses the static external environment recognition information Iprs acquired at the time point t11 in the process Pro5 and the dynamic external world recognition information Iprd acquired at the time point t23 to generate the medium-term trajectory Mt5 (not shown). Is generating.

The use of the static external environment recognition information Iprs at the time point t11 in the lower-layer middle-term orbit generator 72 and the short-term orbit generator 73 is such that the long-term orbit generator 71 generates a long-term orbit Lt2 (not shown) in the process Pro2. When it does, it is stopped and changed to the static external world recognition information Iprs at time t20.

[Summary]
According to the above-described embodiment, the vehicle control device 12 that controls the vehicle 10 that is capable of autonomous driving recognizes and recognizes at least the lane shape of the lane in which the own vehicle 10 travels from the external world information detected by the external world sensor 14. The static external environment recognition information Iprs including the lane shape information Iprsl is generated, and the dynamic external environment recognition information Iprd including the travel suppression source information Iprdc such as an obstacle, a traffic participant, and a traffic light color is generated, and the external environment recognition information is generated. A short-term trajectory as a first trajectory with a short cycle Ts as a first computation cycle in which the computation cycle is relatively short using the external field recognition unit 51 generated as Ipr (Ipr=Iprs+Iprd) and the static external field recognition information Iprs. By using the short-term trajectory generation unit 73 as a first trajectory generation unit that generates St and the external world recognition information Ipr (Ipr=Iprs+Iprd), the short-term trajectory St is used as the second operation cycle longer than the short cycle Ts in the medium-cycle Tm. A medium-term orbit generator 72 that generates a medium-term orbit Mt as a longer second orbit, and a general controller 70 that controls the short-term orbit generator 73 and the medium-term orbit generator 72 are included.

The external world recognition unit 51 performs the recognition process at a cycle of a short cycle Ts (first operation cycle) or less, and the overall control unit 70 uses the static trajectory used by the medium-term trajectory generator 72 (second trajectory generator). When there is a typical external world recognition information Iprs, the static external world recognition information Iprs is used by the short-term trajectory generator 73 (first trajectory generator) to generate the short-term trajectory St (first). Example 1).

In this way, the short-term trajectory St (first trajectory) (first trajectory) is used by causing the short-term trajectory generator 73 (first trajectory generator) to use the static external environment recognition information Iprs used by the medium-term trajectory generator 72 (second trajectory generator). Since the trajectory is generated, the mismatch of the static external environment recognition information Iprs with the medium-term trajectory Mt (second trajectory) used by the short-term trajectory generator 73 (second trajectory generator) is suppressed, and the short-term trajectory is suppressed. A stable trajectory output is performed from the trajectory generation unit 73 (first trajectory generation unit).

More specifically, it is assumed that the short-term orbit generation unit 73 (first orbit generation unit), which is a lower layer, uses the latest external world recognition information Ipr (Ipr=Iprs+Iprd) and the medium-term orbit Mt (second orbit) to optimize the orbit. When the conversion is performed, even if the external world information (environmental information) does not actually change, the medium-term orbit generation unit 72 (second orbit generation unit), which is the upper layer, is caused by the recognition error of the external world recognition unit 51 and the like. The static external world recognition information Iprs different from the used static external world recognition information Iprs will be used.

In this case, the medium-term trajectory Mt (second trajectory) is not an optimal solution for the latest static external environment recognition information Iprs, and the evaluation function for evaluating the short-term trajectory St (first trajectory) {for example, the short-term trajectory When selecting the trajectory point sequence Pj of the short-term trajectory St (first trajectory) from the plurality of trajectory point sequence candidates Pcj generated as St (first trajectory) candidates, basically, the middle-term trajectory of the upper hierarchy Multimodality unnecessary for the function that selects the trajectory point sequence candidate Pcj of the short-term trajectory St (first trajectory) having the highest similarity (small difference) to the trajectory point sequence Pj of Mt (second trajectory)} (originally, There is a possibility that the short-term trajectory St (first trajectory), which is the output trajectory, becomes unstable because the peak point of one point having the highest degree of similarity has multiple peak points.).

In reality, although the recognition error (recognition noise) is added to the lane shape including the white line and the stop line, it is essentially static (does not change or does not move). The external environment recognition information Ipr (the same Iprs is the same, the Iprd is different) that is partially the same (the lane shape information Iprsl and the like) as the static external environment recognition information Iprs referred to by the (second trajectory generation unit) is a short-term trajectory that is a lower layer. By referring to the generation unit 73 (first orbit generation unit), the short-term orbit generation unit 73 (first orbit generation unit), which is a lower layer, and the medium-term orbit Mt (second orbit), which is an upper layer orbit, are referred to. It is possible to suppress the mismatch of the external information (environment map information Iem) and perform stable trajectory output.

As described above, the outside world recognition information Ipr includes, in addition to the static outside world recognition information Iprs including the lane shape information Iprsl, dynamic outside world recognition information Iprd for suppressing the traveling of the own vehicle including the traffic participants. It includes certain travel suppression source information Iprdc. In this case, the integrated control unit 70 causes the short-term orbit generation unit 73 (first orbit generation unit) to use the static external environment recognition information Iprs used by the medium-term orbit generation unit 72 (second orbit generation unit). As for the dynamic external world recognition information Iprd, the latest dynamic external world recognition information Iprd is used.

As described above, the short-term orbit generator 73 as the first orbit generator generates a short period Ts (first operation period) shorter than the medium period Tm (second operation period) of the medium-term orbit generator 72 as the second orbit generator. ), when the short-term orbit St (first orbit) shorter than the medium-term orbit Mt (second orbit) is generated, the dynamic external world recognition information Iprd is recognized in the short cycle Ts (first operation cycle) or less. Since it is generated by referring to the latest external world recognition information Iprd, a short-term trajectory St (first trajectory) that does not reduce the responsiveness of the vehicle 10 to traffic participants and the like can be generated.

As described above, the outside world recognition information Ipr includes, in addition to the lane shape information Iprsl, the sign/marking information Iprsm for restricting the traveling of the vehicle 10, and the overall control unit 70 determines the sign/marking information. Iprsm is also used by the medium-term orbit generator 72 (second orbit generator), and the sign/marking information Iprsm used by the medium-term orbit generator 72 is used by the short-term orbit generator 73 (first orbit generator).

For example, the sign/marking information Iprsm used by the medium-term orbit generator 72 (second orbit generator) for restricting traveling of a sign indicating the maximum speed, a stop line, or the like is set to the short-term orbit generator 73 (first orbit generator). ), the medium-term trajectory Mt (second trajectory) that performs smooth transition of the velocity vs of the vehicle 10 generated by the medium-term trajectory generator 72 (second trajectory generator) and smooth stop of the vehicle 10 is used for a short period. It can be used by the trajectory generator 73 (first trajectory generator).

In this case, with reference to the static external environment recognition information Iprs in the long cycle Tl as the third calculation cycle, which is longer than the middle cycle Tm (second calculation cycle) controlled by the central control unit 70, the middle-term trajectory Mt is referred to. When the long-term orbit generating unit 71 as the third orbit generating unit that generates the long-term orbit Lt as the longer third orbit is included, the overall control unit 70, when starting the automatic operation, the short-term orbit generating unit 73 (first The orbit generation unit), the medium-term orbit generation unit 72 (second orbit generation unit), and the long-term orbit generation unit 71 (third orbit generation unit) simultaneously start generation of each orbit, while the medium-term orbit Mt (second orbit). Before the trajectory is generated, the short-term trajectory generator 73 (first trajectory generator) is caused to generate the short-term trajectory St (first trajectory) by referring to the latest external world recognition information Ipr, and the medium-term trajectory Mt. When the (second orbit) is generated, the medium-term orbit Mt (second orbit) and the medium-term orbit generator 72 (second orbit generation unit) are compared with the short-term orbit generator 73 (first orbit generator). When the short-term trajectory St (first trajectory) is generated by referring to the static external environment recognition information Iprs including the referenced lane shape, and the long-term trajectory Lt (third trajectory) is generated, the short-term trajectory generation unit 73 On the other hand, the short-term trajectory St is generated by using the external environment recognition information Ipr including the lane shape information Iprsl used by the long-term trajectory Lt (the third trajectory) and the long-term trajectory generator 71 (the third trajectory generator). May be.

When the automatic operation is turned on, the short-term trajectory generation unit 73, the medium-term trajectory generation unit 72, and the long-term trajectory generation unit 71 first use the short-term trajectory generation unit 73 (first trajectory) at the shortest computation cycle Toc. The vehicle 10 is controlled by the short-term trajectory St generated by the generation unit), and then the medium-term trajectory Mt( generated by the medium-term trajectory generation unit 72 (second trajectory generation unit) at the next shortest computation cycle 3×Toc. The vehicle 10 is controlled by the short-term trajectory St (first trajectory) considering the second trajectory), and then the medium-term considering the long-term trajectory Lt generated by the long-term trajectory generation unit 71 at the longest calculation cycle 9×Toc. The vehicle 10 is controlled by the short-term track St generated in consideration of the track Mt. Therefore, it is possible to immediately start the automatic driving based on the short-term track St, and to gradually and gradually shift to the automatic driving in consideration of the medium-term track Mt and the long-term track Lt in which the riding comfort and comfort are improved. In this case, the upper-layer orbits (the middle-term orbits) referred to by the lower-level orbit generators (the medium-term orbit generator 72 for the long-term orbit generator 71, and the short-term orbit generators 73 for the long-term orbit medium generator 71 and the medium-term orbit generator 72) are referred to. It is possible to suppress the mismatch of the external environment information (environment map information Iem) with the long-term trajectory Lt with respect to the trajectory Mt, and the medium-term trajectory Mt and the long-term trajectory Lt with respect to the short-term trajectory St, and perform stable trajectory output.

The external environment recognition information Ipr includes travel suppression source information Iprdc that suppresses travel of the vehicle 10, in addition to the lane shape information Iprsl or the lane shape information Iprsl and the sign/marking information Iprsm. The control unit 70 does not use (reference) the travel suppression source information Iprdc to the long-term trajectory generation unit 71, but only refers to (uses) the short-term trajectory generation unit 73 (first trajectory generation unit).

In the long-term track Lt as the third track which is a relatively long track, the riding comfort (smoothness of changes in vehicle behavior) is emphasized, and the medium-term track Mt and the relative track as the second track, which is a relatively middle track, Since the adaptability and responsiveness to the recognized external environment are emphasized in the short-term trajectory St as the first trajectory, which is a relatively short trajectory, the traveling of the host vehicle 10 is performed when the short-term trajectory St and the medium-term trajectory Mt are generated. By referring to the travel restraint source information Iprdc, the vehicle 10 is controlled by the short-term trajectory St that indirectly refers to the long-term trajectory Lt generated by referring to the lane shape while ensuring adaptability and responsiveness to the external environment. By controlling the, it is possible to perform automatic driving in which ride comfort (comfort) is taken into consideration.

The lane shape information Iprsl is information recognized from lane regulations such as the lane L or curb Es and guardrail Gr provided on the road surface, and the travel restraint source information Iprdc is an obstacle, a traffic participant, or a traffic light color. It is information including.

The long-term track generation unit 71 (third track generation unit) can generate a track (long-term track Lt) that emphasizes ride comfort based on information recognized from the lane regulation provided on the road surface, and the short-term track generation unit 73 ( The first trajectory generation unit) and the medium-term trajectory generation unit 72 (second trajectory generation unit) are trajectory (short-term trajectory St) that emphasizes adaptability/responsiveness based on information including obstacles, traffic participants, or traffic light colors. , The mid-term trajectory Mt ) can be generated.

The lane standard includes a lane deviation preventing member such as a lane mark Lm and/or a curb Es, a guardrail Gr, or the like.

The outside world recognition unit 51 can recognize the lane shape of the lane L on which the vehicle 10 travels from the lane departure prevention member such as the lane mark Lm or the curb Es/guard rail Gr.

The present invention is not limited to the above-described embodiment, and it goes without saying that various configurations can be adopted based on the contents of this specification.

Claims (6)

A vehicle control device (12) for controlling a vehicle (10) capable of autonomous driving, comprising:
At least the lane shape of the lane in which the vehicle is traveling is recognized from the outside world information detected by the outside world sensor (14) that detects the outside world information, and the static outside world recognition information (Iprs) including the recognized lane shape information (Iprsl) is recognized. ) And an external environment recognition unit (51) that generates dynamic external environment recognition information (Iprd) including travel suppression source information (Iprdc) that suppresses the travel of the vehicle including the traffic participant,
A first trajectory generation unit that generates a first trajectory in a first calculation cycle by using the outside world recognition information;
A second trajectory generation unit that generates a second trajectory longer than the first trajectory in a second computation cycle longer than the first computation cycle using the external world recognition information;
An overall control unit (70) for controlling the first trajectory generation unit and the second trajectory generation unit;
Equipped with
The surroundings recognition part (51), which performs the recognition processing in a cycle of less than the first operation cycle,
The integrated control unit (70)
In forming the latest the first track by said first track generation unit, to use the latest dynamic the surroundings recognition information (IPRD), and said second trajectory generation unit is the last of the second track When there is the static external environment recognition information (Iprs) before the second operation cycle used when generating, the static external environment recognition information (Iprs) before the second operation cycle and the latest The vehicle control device (12) , wherein the latest first track is generated by referring to the second track generated by the above . The vehicle control device (12) according to claim 1,
The outside world recognition information includes, in addition to the lane shape information (Iprsl), static sign outside world recognition information (Iprs) that restricts the traveling of the own vehicle (10) (Iprsm). Is something
The integrated control unit (70)
The second trajectory generation unit is caused to use the marking/marking information (Iprsm), and the first trajectory is used for the marking/marking information (Iprsm) used by the second trajectory generating unit before the second calculation cycle. A vehicle control device (12) characterized by being used by a generation unit. The vehicle control device (12) according to claim 1 or 2 ,
A third trajectory generation unit that generates a third trajectory longer than the second trajectory by using the external world recognition information in a third computation cycle that is longer than the second computation cycle and that is controlled by the central control unit (70). Further has
The integrated control unit (70)
When starting the automatic operation, the first trajectory generation unit, the second trajectory generation unit, and the third trajectory generation unit simultaneously start generation of each trajectory,
Before the latest second trajectory is generated, the first trajectory generation unit is caused to use the latest external world recognition information to generate the latest first trajectory,
When the most recent second trajectory is generated, when the most recently generated second trajectory and the second trajectory generation section generate the most recent second trajectory with respect to the first trajectory generation section. Using the static external environment recognition information before the second operation cycle including the used lane shape, the latest first trajectory is generated,
When the most recent of the third trajectory is generated, relative to the first trajectory generating section, the third track and the third track generation unit generated most recently utilized in generating the third track A vehicle control device (12), characterized in that the latest first track is generated by using the external environment recognition information including the lane shape information (Iprsl) before the third calculation cycle . The vehicle control device (12) according to claim 3 ,
The outside world recognition information further includes travel restraint source information (Iprdc) for restraining the traveling of the own vehicle (10),
The integrated control unit (70)
The vehicle control device (12), wherein the traveling suppression source information (Iprdc) is not used by the third track generation unit but is used by the first and second track generation units. The vehicle control device (12) according to claim 4 ,
The lane shape information (Iprsl) is information recognized from a lane prescription provided on a road surface, and the travel restraint source information (Iprdc) is information including an obstacle, a traffic participant, or a traffic light color. A vehicle control device (12) characterized by the following. The vehicle control device (12) according to claim 5 ,
The vehicle control device (12), wherein the lane regulation includes a lane mark or a lane departure prevention member.

Images